“We’re ready to push the button,” Jeroen, my co-founder and VP Engineering, said to me.
“You’ve signed off on all the design changes that needed to be made?” I asked Jeroen.
“Yes. Sid has done everything on the list,” Jeroen said.
“Okay, then let’s do it.”
And with my agreement, we just spent $180,000 “taping out” our first product.
I’d approved tape outs a million times before, so the process was nothing new to me. However, this was the first time, as founding CEO of my company, that I’d agreed to “push the button.”
By modern standards, spending $180,000 to tape out (this means sending the electronic file to your fabrication facility, so they have the instructions needed to make an integrated circuit) is a relatively small number. Cutting edge technologies, not the analog trailing edge technology we were using, could cost over $1 million to tape out.
Mistakes in some industries will put you out of business.
It’s like Jack Nicholson said in A Few Good Men, “We follow orders or people die.”
Our whole business model depended upon us having at least fifty percent of our products have “first silicon” success. That meant we could introduce the first article of a new device as a production product.
A 50% success rate was what a well-run analog IC company could reliably achieve. Getting to this type of success level meant having a rigorous peer design review process.
Rigor doesn’t have to come at the expense of speed. In fact, you can move faster with rigor.
I had the benefit of working at two companies in the same industry, one after the other. The companies were diametrically opposites.
The first company, Maxim Integrated Products, had an extremely disciplined engineering organization. It’s where the bulk of our engineers came from.
The last division I ran at Maxim was its wireline communications division. We averaged getting a product introduced to market, from start to finish, in 12 months.
Now, let’s take the second company, Micrel. The division I took over at Micrel was its wireline communications division. So the comparison is apples to apples.
Engineering for Micrel's wireline communications division was run by “The Michelin Man” (so named because he liked to fold his arms around himself as he pondered your question, making him look like The Michelin Man).
The Michelin Man’s team averaged around 3 years to get similarly complex products out the door.
Oh, the Michelin Man liked to move fast. But, the difference between the companies came down to two things:
- The quality of the engineers
- The rigor of the design review process
The Maxim team was disciplined and the Micrel team wasn’t disciplined. The Michelin Man would tape out a device with no peer review.
Then, The Michelin Man would make matters worse when the product came back from the fab and, inevitably, fail. He would rush to make changes to the product without even simulating the results.
In other words, The Michelin Man was blindly making changes by gut instinct rather than the data. But he was moving fast and breaking things, that’s for sure.
It was a perpetual doom loop.
The Michelin Man paid for his sloppiness with his job.
You can be disciplined and move fast.
In an industry where the cost of failure is small and you can iterate quickly, then move fast and break things sounds great, doesn’t it. However, in an industry where breaking things has a huge downside, then you have to move in a fast and disciplined way.